R Notebook

devtools::install_github("lme4/lme4",dependencies=TRUE)

Downloading GitHub repo lme4/lme4@master
from URL https://api.github.com/repos/lme4/lme4/zipball/master
Installation failed: cannot open file 'C:/Users/Fabio/Documents/R/win-library/3.4/lme4/R/sysdata.rdb': No such file or directory

Load in Data

arab4 <- read_spss("data/arab4.sav")
arab3 <- read_spss("data/arab3.sav")
# arab4 <- get(load(url("https://github.com/favstats/GodlyGovernance/raw/master/data/arab4.Rdata")))

Inspect Data

#b_select <- binoculaR::binoculaR(arab4)
#dput(b_select$var_codes)
#index <- c("country","wt" ,"sample" ,"a1", "q1", "q2", "q13", "q5182", 
#"q5184", "q6013", "q6014", "q6018", "q60118", "q605a", "q6061", "q6062", 
#"q6063", "q6064", "q6071", "q6074", "q6076", "q707", "q708", 
#"q7114", "q1001", "q1002", "q1003", "t1003", "q1004", "q1005", 
#"q609", "q6101", "q6106", "q1012", "q1012a", "q1016")
#
#arab4 %<>%
#  select(index)
#
#arab4
#view_df2(arab3, hide.progress = T)
# save(arab4, file = "data/arab4.Rdata")

Filter Data

arab4 %<>%
  filter(q1012 == 1) %>%#only Muslims
  mutate(sample = ifelse(is.na(sample) | sample == 1, 1, 2)) %>% 
  filter(sample != 2) #only non-refugees
arab3 %<>%
  filter(q1012 == 1) #only Muslims
arab3

arab4

Recode Data

Arab3

arab3 %<>% 
  mutate(cntry = sjmisc::to_label(country)) %>% 
  mutate(region = sjmisc::to_label(a1)) %>% 
  mutate(governorate = sjmisc::to_label(q1)) %>% 
  mutate(year = lubridate::year(date))
  # Dependent Variable
arab3 %<>% 
  mutate(islamistparties1 = ifelse(q518a2 > 5, NA, 5 - q518a2)) %>%
  mutate(islamistparties2 = ifelse(q518b2 > 5, NA, 5 - q518b2)) %>% 
  mutate(islamistparties = case_when(
        islamistparties1 == 1 ~ 1,
        islamistparties1 == 2 ~ 2,
        islamistparties1 == 3 ~ 3,
        islamistparties1 == 4 ~ 4,
        islamistparties2 == 1 ~ 1,
        islamistparties2 == 2 ~ 2,
        islamistparties2 == 3 ~ 3,
        islamistparties2 == 4 ~ 4)
    ) %>%  
  mutate(islamistgov = ifelse(q5184 > 5, NA, 5 - q5184)) %>% 
  mutate(religinterfere = ifelse(q6061 > 5, NA, q6061)) %>% 
  mutate(religleaders = ifelse(q6062 > 5, NA, 5 - q6062)) %>% 
  mutate(religleadersinfl = ifelse(q6063 > 5, NA, 5 - q6063)) %>%
  mutate(seperation = ifelse(q6064 > 5, NA, q6064)) %>% 
  mutate(religparty = Recode(q605a, "1 = 1;
                                     2 = 1;
                                     3 = 0;
                                     4 = 0;
                                     5 = 0;
                                    8 = 0;
                                    9 = NA")) %>% 
  mutate(religparty2 = Recode(q605a, "1 = 5;
                                      2 = 4;
                                      3 = 2;
                                      4 = 1;
                                      5 = 3;
                                     8 = NA;
                                     9 = NA"))
ifelse4cat_rec <- function(variable) {
  recoded <- ifelse(variable == 0 | variable > 5, NA, 5 - variable)
  return(recoded)
}
arab3 %<>% 
  mutate(female = ifelse(sex == 2, 1, 0)) %>% 
  mutate(work = ifelse(q1004 == 0 | q1004 > 5, NA, abs(q1004 - 2))) %>% 
  mutate(income = ifelse4cat_rec(q1016)) %>% 
  mutate(age = ifelse(q1001 == 0 | q1001 == 9999, NA, q1001)) %>% 
  mutate(educ = case_when(
        q1003 == 0 ~ NA_real_,
        q1003 == 99 ~ NA_real_,
        q1003 == 5 ~ 4,
        q1003 == 6 ~ 5,
        q1003 == 7 ~ 6,
        q1003yem == 0 ~ NA_real_,
        q1003yem == 99 ~ NA_real_,
        q1003yem == 4 ~ 3,
        q1003yem == 5 ~ 4,
        q1003yem == 6 ~ 5,
        q1003yem == 7 ~ 5,
        q1003yem == 8 ~ 6,
        q1003t == 0 ~ NA_real_,
        q1003t == 99 ~ NA_real_,
        q1003t == 1 ~ 1,
        q1003t == 2 ~ 2,
        q1003t == 3 ~ 3,
        q1003t == 4 ~ 4,
        q1003t == 5 ~ 5,
        q1003t == 6 ~ 6,
    TRUE ~ as.numeric(q1003))
    ) %>% 
  mutate(globalism = ifelse(q701b == 0 | q701b > 5, NA, q701b)) %>% 
  mutate(pray = ifelse(q6101 == 0 | q6101 > 5, NA, 6 - q6101)) %>% 
  mutate(quran = ifelse(q6106 == 0 | q6106 > 5, NA, 6 - q6106)) %>%
  mutate(womanwork = ifelse(q6012 == 0 | q6012 > 5, NA, q6012)) %>% 
  mutate(womenleader = ifelse4cat_rec(q6013)) %>% 
  mutate(womeneduc = ifelse4cat_rec(q6014)) %>% 
  mutate(nodemoc = ifelse(q6071 == 0 | q6071 > 5, NA, q6071)) %>% 
  mutate(genderapartuni = ifelse4cat_rec(q6074)) %>% 
  mutate(coverup = ifelse4cat_rec(q6076)) %>% 
  select(cntry, year, region, governorate, islamistparties , islamistgov, religinterfere, religleaders, religleadersinfl, seperation, religparty, religparty2, female, work, income,  age, educ, globalism, pray, quran, womanwork, womenleader, womeneduc, nodemoc, genderapartuni, coverup)

Arab4

table(arab4$country)

   1    5    8   10   13   15   21 
1199 1146 1172  615 1199 1177 1200 

arab4 %<>% 
  mutate(cntry = sjmisc::to_label(country)) %>% 
  mutate(region = sjmisc::to_label(a1)) %>% 
  mutate(governorate = sjmisc::to_label(q1)) %>% 
  mutate(year = 2016)#%>%
#  mutate(district = sjmisc::to_label(q2)) 
# Dependent Variable
arab4 %<>% 
  mutate(islamistparties = ifelse(q5182 > 5, NA, 5 - q5182)) %>% 
  mutate(islamistgov = ifelse(q5184 > 5, NA, 5 - q5184)) %>% 
  mutate(religinterfere = ifelse(q6061 > 5, NA, q6061)) %>% 
  mutate(religleaders = ifelse(q6062 > 5, NA, 5 - q6062)) %>% 
  mutate(religleadersinfl = ifelse(q6063 > 5, NA, 5 - q6063)) %>%
  mutate(seperation = ifelse(q6064 > 5, NA, q6064)) %>% 
  mutate(religparty = Recode(q605a, "1 = 1;
                                     2 = 1;
                                     3 = 0;
                                     4 = 0;
                                     5 = 0;
                                    98 = 0;
                                    99 = NA")) %>% 
  mutate(religparty2 = Recode(q605a, "1 = 5;
                                      2 = 4;
                                      3 = 2;
                                      4 = 1;
                                      5 = 3;
                                     98 = NA;
                                     99 = NA"))
arab4 %<>% 
  mutate(female = ifelse(q1002 == 2, 1, 0)) %>% 
  mutate(work = ifelse(q1004 == 0 | q1004 > 5, NA, abs(q1004 - 2))) %>% 
  mutate(income = ifelse4cat_rec(q1016)) %>% 
  mutate(age = ifelse(q1001 == 0 | q1001 == 9999, NA, q1001)) %>% 
  mutate(educ = case_when(
        q1003 == 0 ~ NA_real_,
        q1003 == 98 ~ NA_real_,
        q1003 == 99 ~ NA_real_,
        q1003 == 5 ~ 4,
        q1003 == 6 ~ 5,
        q1003 == 7 ~ 6,
        t1003 == 0 ~ NA_real_,
        t1003 == 98 ~ NA_real_,
        t1003 == 99 ~ NA_real_,
        t1003 == 3 ~ 3,
        t1003 == 4 ~ 4,
        t1003 == 5 ~ 5,
        t1003 == 6 ~ 6,
    TRUE ~ as.numeric(q1003))
    ) %>% 
  mutate(globalism = ifelse(q701b == 0 | q701b > 5, NA, q701b)) %>% 
  mutate(pray = ifelse(q6101 == 0 | q6101 > 5, NA, 6 - q6101)) %>% 
  mutate(quran = ifelse(q6106 == 0 | q6106 > 5, NA, 6 - q6106)) %>% 
  mutate(womanwork = ifelse(q6012 == 0 | q6012 > 5, NA, q6012)) %>% 
  mutate(womenleader = ifelse4cat_rec(q6013)) %>% 
  mutate(womeneduc = ifelse4cat_rec(q6014)) %>% 
  mutate(nodemoc = ifelse(q6071 == 0 | q6071 > 5, NA, q6071)) %>% 
  mutate(genderapartuni = ifelse4cat_rec(q6074)) %>% 
  mutate(coverup = ifelse4cat_rec(q6076)) %>% 
  select(cntry, year, region, governorate , islamistparties , islamistgov, religinterfere, religleaders, religleadersinfl, seperation, religparty, religparty2, female, work, income,  age, educ, globalism, pray, quran, womanwork, womenleader, womeneduc, nodemoc, genderapartuni, coverup)

Merging

Factor Analysis

f31 <- arab %>% 
  select(islamistparties, islamistgov, 
         religleaders, religleadersinfl) %>% 
  fa(1, rotate = "promax",   
        fm = "pa",
        scores = "regression")           
fa.diagram(f31)                                  

f31              

Factor Analysis using method =  pa
Call: fa(r = ., nfactors = 1, rotate = "promax", scores = "regression", 
    fm = "pa")
Standardized loadings (pattern matrix) based upon correlation matrix
                  PA1   h2   u2 com
islamistparties  0.55 0.30 0.70   1
islamistgov      0.54 0.29 0.71   1
religleaders     0.71 0.50 0.50   1
religleadersinfl 0.63 0.40 0.60   1

                PA1
SS loadings    1.50
Proportion Var 0.38

Mean item complexity =  1
Test of the hypothesis that 1 factor is sufficient.

The degrees of freedom for the null model are  6  and the objective function was  0.75 with Chi Square of  16491.97
The degrees of freedom for the model are 2  and the objective function was  0.09 

The root mean square of the residuals (RMSR) is  0.08 
The df corrected root mean square of the residuals is  0.14 

The harmonic number of observations is  19654 with the empirical chi square  1517.15  with prob <  0 
The total number of observations was  21915  with Likelihood Chi Square =  1941.79  with prob <  0 

Tucker Lewis Index of factoring reliability =  0.647
RMSEA index =  0.21  and the 90 % confidence intervals are  0.203 0.218
BIC =  1921.8
Fit based upon off diagonal values = 0.96
Measures of factor score adequacy             
                                                   PA1
Correlation of (regression) scores with factors   0.85
Multiple R square of scores with factors          0.72
Minimum correlation of possible factor scores     0.43

arab %>% 
  select(islamistparties, islamistgov, 
         religleaders, religleadersinfl) %>% 
  alpha() 

Reliability analysis   
Call: alpha(x = .)

  raw_alpha std.alpha G6(smc) average_r S/N    ase mean   sd
       0.7       0.7    0.66      0.37 2.3 0.0034    2 0.71

 lower alpha upper     95% confidence boundaries
0.69 0.7 0.71 

 Reliability if an item is dropped:
                 raw_alpha std.alpha G6(smc) average_r S/N alpha se
islamistparties       0.65      0.65    0.58      0.39 1.9   0.0042
islamistgov           0.65      0.66    0.58      0.39 1.9   0.0041
religleaders          0.60      0.60    0.51      0.33 1.5   0.0046
religleadersinfl      0.64      0.64    0.54      0.37 1.7   0.0042

 Item statistics 
                     n raw.r std.r r.cor r.drop mean   sd
islamistparties  20198  0.72  0.71  0.55   0.46  1.7 0.95
islamistgov      20135  0.73  0.71  0.54   0.46  1.7 1.02
religleaders     20431  0.75  0.76  0.66   0.53  2.3 0.92
religleadersinfl 20524  0.72  0.73  0.60   0.48  2.2 0.89

Non missing response frequency for each item
                    1    2    3    4 5 miss
islamistparties  0.59 0.19 0.15 0.07 0 0.08
islamistgov      0.59 0.17 0.14 0.10 0 0.08
religleaders     0.21 0.38 0.31 0.10 0 0.07
religleadersinfl 0.24 0.42 0.26 0.08 0 0.06

arab %>% 
  select(islamistparties, islamistgov, 
         religleaders, religleadersinfl) %>% 
  KMO() 

Kaiser-Meyer-Olkin factor adequacy
Call: KMO(r = .)
Overall MSA =  0.68
MSA for each item = 
 islamistparties      islamistgov     religleaders religleadersinfl 
            0.71             0.72             0.66             0.66 

arab4 %>% 
  select(islamistparties, islamistgov, 
         religleaders, religleadersinfl,
         religparty2, seperation,
         religinterfere) %>% 
  na.omit() %>% 
  cor()  

                 islamistparties islamistgov religleaders religleadersinfl
islamistparties        1.0000000   0.3760955    0.3274069        0.2595854
islamistgov            0.3760955   1.0000000    0.3180997        0.3059820
religleaders           0.3274069   0.3180997    1.0000000        0.5628358
religleadersinfl       0.2595854   0.3059820    0.5628358        1.0000000
religparty2            0.2625742   0.2409189    0.3812004        0.3120300
seperation             0.1409401   0.2186956    0.3190967        0.2944263
religinterfere         0.1393481   0.1283822    0.2070613        0.2539891
                 religparty2 seperation religinterfere
islamistparties    0.2625742  0.1409401      0.1393481
islamistgov        0.2409189  0.2186956      0.1283822
religleaders       0.3812004  0.3190967      0.2070613
religleadersinfl   0.3120300  0.2944263      0.2539891
religparty2        1.0000000  0.2819491      0.1671423
seperation         0.2819491  1.0000000      0.2725233
religinterfere     0.1671423  0.2725233      1.0000000

arab <- predict(f31, arab %>% 
  select(islamistparties, islamistgov, 
         religleaders, religleadersinfl)) %>% 
  data.frame() %>% 
  cbind(arab) %>% 
  mutate(islamism = PA1) #%>% 
#  ggplot(aes(islamism)) +
#  geom_histogram() +
#  facet_wrap(~cntry, scales = "free")
#sjPlot::view_df(arab, show.frq = T, show.prc = T)

Creating indices

Regression

model0 <- arab_reg %>% 
  lme4::lmer(islamism ~ 1 + (1|cntry/year), data = .)
icc(model0)

Linear mixed model
 Family: gaussian (identity)
Formula: islamism ~ 1 + (1 | cntry/year)

  ICC (year:cntry): 0.036909
       ICC (cntry): 0.119094

model1 <- lme4::lmer(islamism ~ female + work + income + age + educ + globalism + personalpiety + patriarchalvalues + liberalislam + (1|cntry/year), data = arab_reg)
texreg::screenreg(model1)

==========================================
                             Model 1      
------------------------------------------
(Intercept)                      33.74 ***
                                 (2.05)   
female                            1.92 ***
                                 (0.38)   
work                             -1.27 ***
                                 (0.38)   
income                            0.29    
                                 (0.18)   
age                              -1.05 ***
                                 (0.19)   
educ                             -0.78 ***
                                 (0.20)   
globalism                         1.67 ***
                                 (0.18)   
personalpiety                     2.89 ***
                                 (0.18)   
patriarchalvalues                 3.84 ***
                                 (0.19)   
liberalislam                     -4.90 ***
                                 (0.19)   
------------------------------------------
AIC                          118421.07    
BIC                          118518.71    
Log Likelihood               -59197.54    
Num. obs.                     13499       
Num. groups: year:cntry          19       
Num. groups: cntry               12       
Var: year:cntry (Intercept)      13.82    
Var: cntry (Intercept)           39.26    
Var: Residual                   375.00    
==========================================
*** p < 0.001, ** p < 0.01, * p < 0.05

plot_model(model1, type = "re", sort.est = T, show.values = T, show.p = T, value.offset = 0.5)

Sorting each group of random effects ('sort.all') is not possible when 'facets = TRUE'.
Sorting each group of random effects ('sort.all') is not possible when 'facets = TRUE'.

[[1]]

[[2]]

#plot_model(model1, type = "std", sort.est = T, show.values = T, show.p = T, value.offset = 0.5)
sjp.lmer(model1, type = "fe.std", sort.est = T, show.values = T, show.p = T, y.offset = 0.4, p.kr = FALSE)

Computing p-values via Wald-statistics approximation (treating t as Wald z).

plot_model(model1, terms = c("liberalislam", "personalpiety"), type = "pred")

Note: uncertainty of the random effects parameters are not taken into account for confidence intervals.

icc(model1)

Linear mixed model
 Family: gaussian (identity)
Formula: islamism ~ female + work + income + age + educ + globalism + personalpiety + patriarchalvalues + liberalislam + (1 | cntry/year)

  ICC (year:cntry): 0.032275
       ICC (cntry): 0.091703

model2 <- lme4::lmer(islamism ~ female + work + income + age + educ + globalism + personalpiety + patriarchalvalues + liberalislam + liberalislam*personalpiety + (1|cntry/year), data = arab_reg)
texreg::screenreg(model2)

==========================================
                             Model 1      
------------------------------------------
(Intercept)                      33.66 ***
                                 (2.05)   
female                            1.96 ***
                                 (0.38)   
work                             -1.25 ** 
                                 (0.38)   
income                            0.30    
                                 (0.18)   
age                              -1.05 ***
                                 (0.19)   
educ                             -0.79 ***
                                 (0.20)   
globalism                         1.69 ***
                                 (0.18)   
personalpiety                     2.96 ***
                                 (0.18)   
patriarchalvalues                 3.85 ***
                                 (0.19)   
liberalislam                     -4.87 ***
                                 (0.19)   
personalpiety:liberalislam       -0.48 ** 
                                 (0.16)   
------------------------------------------
AIC                          118416.31    
BIC                          118521.45    
Log Likelihood               -59194.15    
Num. obs.                     13499       
Num. groups: year:cntry          19       
Num. groups: cntry               12       
Var: year:cntry (Intercept)      14.23    
Var: cntry (Intercept)           38.96    
Var: Residual                   374.79    
==========================================
*** p < 0.001, ** p < 0.01, * p < 0.05

plot_model(model2, sort.est = T, show.values = T, show.p = T, value.offset = 0.5)

Computing p-values via Wald-statistics approximation (treating t as Wald z).

sjp.int(model2, p.kr = FALSE)

`sjp.int()` will become deprecated in the future. Please use `plot_model()` instead.

sjp.lmer(model2, type = "fe.std", sort.est = T, show.values = T, show.p = T, y.offset = 0.4, p.kr = FALSE)

Computing p-values via Wald-statistics approximation (treating t as Wald z).

#plot_model(model2, type = "int")
vif.mer <- function (fit) {
  ## adapted from rms::vif
  
  v <- vcov(fit)
  nam <- names(fixef(fit))
  
  ## exclude intercepts
  ns <- sum(1 * (nam == "Intercept" | nam == "(Intercept)"))
  if (ns > 0) {
    v <- v[-(1:ns), -(1:ns), drop = FALSE]
    nam <- nam[-(1:ns)]
  }
  
  d <- diag(v)^0.5
  v <- diag(solve(v/(d %o% d)))
  names(v) <- nam
  v
}
vif.mer(model2)

                    female                       work 
                  1.313155                   1.270084 
                    income                        age 
                  1.062632                   1.074094 
                      educ                  globalism 
                  1.155690                   1.052950 
             personalpiety          patriarchalvalues 
                  1.076342                   1.162692 
              liberalislam personalpiety:liberalislam 
                  1.109287                   1.026315 

model1 <- arab_reg %>% 
  lme4::lmer(islamism ~ female + work + income + age + educ + globalism + pray + quran + womanwork + womenleader + womeneduc + nodemoc + genderapartuni + coverup + (1|cntry/year), data = .)
texreg::screenreg(model1)

==========================================
                             Model 1      
------------------------------------------
(Intercept)                      33.38 ***
                                 (2.54)   
female                            1.89 ***
                                 (0.38)   
work                             -1.31 ***
                                 (0.38)   
income                            0.28    
                                 (0.18)   
age                              -1.02 ***
                                 (0.19)   
educ                             -0.74 ***
                                 (0.20)   
globalism                         1.63 ***
                                 (0.18)   
pray                              1.31 ***
                                 (0.20)   
quran                             2.05 ***
                                 (0.19)   
womanwork                         0.78 ***
                                 (0.22)   
womenleader                       2.40 ***
                                 (0.18)   
womeneduc                         2.14 ***
                                 (0.20)   
nodemoc                          -3.63 ***
                                 (0.21)   
genderapartuni                   -4.56 ***
                                 (0.22)   
coverup                          -0.82 ***
                                 (0.18)   
------------------------------------------
AIC                          118239.29    
BIC                          118374.48    
Log Likelihood               -59101.64    
Num. obs.                     13499       
Num. groups: year:cntry          19       
Num. groups: cntry               12       
Var: year:cntry (Intercept)      14.12    
Var: cntry (Intercept)           37.00    
Var: Residual                   369.66    
==========================================
*** p < 0.001, ** p < 0.01, * p < 0.05

plot_model(model1, sort.est = T, show.values = T, show.p = T, value.offset = 0.5)

Computing p-values via Wald-statistics approximation (treating t as Wald z).

plot_model(model1, terms = c("nodemoc", "female"), type = "pred")

Note: uncertainty of the random effects parameters are not taken into account for confidence intervals.

model2 <- arab_reg %>% 
  lme4::lmer(islamism ~ female + work + income + age + educ + globalism + pray + quran + womanwork + womenleader + womeneduc + nodemoc + genderapartuni + coverup + nodemoc*quran + (1|cntry/year), data = .)
texreg::screenreg(model2)

==========================================
                             Model 1      
------------------------------------------
(Intercept)                      35.34 ***
                                 (3.22)   
female                            1.89 ***
                                 (0.38)   
work                             -1.31 ***
                                 (0.38)   
income                            0.28    
                                 (0.18)   
age                              -1.02 ***
                                 (0.19)   
educ                             -0.74 ***
                                 (0.20)   
globalism                         1.63 ***
                                 (0.18)   
pray                              1.32 ***
                                 (0.20)   
quran                             1.53 ** 
                                 (0.56)   
womanwork                         0.79 ***
                                 (0.22)   
womenleader                       2.40 ***
                                 (0.18)   
womeneduc                         2.14 ***
                                 (0.20)   
nodemoc                          -4.32 ***
                                 (0.72)   
genderapartuni                   -4.55 ***
                                 (0.22)   
coverup                          -0.82 ***
                                 (0.18)   
quran:nodemoc                     0.18    
                                 (0.18)   
------------------------------------------
AIC                          118241.91    
BIC                          118384.60    
Log Likelihood               -59101.95    
Num. obs.                     13499       
Num. groups: year:cntry          19       
Num. groups: cntry               12       
Var: year:cntry (Intercept)      14.13    
Var: cntry (Intercept)           37.02    
Var: Residual                   369.66    
==========================================
*** p < 0.001, ** p < 0.01, * p < 0.05

view_df2 <- function (x, weight.by = NULL, altr.row.col = TRUE, show.id = TRUE, 
  show.type = FALSE, show.values = TRUE, show.string.values = FALSE, 
  show.labels = TRUE, show.frq = FALSE, show.prc = FALSE, 
  show.wtd.frq = FALSE, show.wtd.prc = FALSE, show.na = FALSE, 
  max.len = 15, sort.by.name = FALSE, wrap.labels = 50, hide.progress = FALSE, 
  CSS = NULL, encoding = NULL, file = NULL, use.viewer = TRUE, 
  no.output = FALSE, remove.spaces = TRUE) 
{
  get.encoding <- function(encoding, data = NULL) {
  if (is.null(encoding)) {
    if (!is.null(data) && is.data.frame(data)) {
      # get variable label
      labs <- sjlabelled::get_label(data[[1]])
      # check if vectors of data frame have
      # any valid label. else, default to utf-8
      if (!is.null(labs) && is.character(labs))
        encoding <- Encoding(sjlabelled::get_label(data[[1]]))
      else
        encoding <- "UTF-8"
      # unknown encoding? default to utf-8
      if (encoding == "unknown") encoding <- "UTF-8"
    } else if (.Platform$OS.type == "unix")
      encoding <- "UTF-8"
    else
      encoding <- "Windows-1252"
  }
  return(encoding)
}

  has_value_labels <- function(x) {
  !(is.null(attr(x, "labels", exact = T)) && is.null(attr(x, "value.labels", exact = T)))
}

  sju.rmspc <- function(html.table) {
  cleaned <- gsub("      <", "<", html.table, fixed = TRUE, useBytes = TRUE)
  cleaned <- gsub("    <", "<", cleaned, fixed = TRUE, useBytes = TRUE)
  cleaned <- gsub("  <", "<", cleaned, fixed = TRUE, useBytes = TRUE)
  return(cleaned)
  }
  
  frq.value <- function(index, x, df.val, weights = NULL) {
  valstring <- ""
  # check if we have a valid index
  if (index <= ncol(x) && !is.null(df.val[[index]])) {
    # do we have weights?
    if (!is.null(weights))
      variab <- sjstats::weight(x[[index]], weights)
    else
      variab <- x[[index]]
    # create frequency table. same function as for
    # sjt.frq and sjp.frq
    ftab <- create.frq.df(variab, 20)$mydat$frq
    # remove last value, which is N for NA
    if (length(ftab) == 1 && is.na(ftab)) {
      valstring <- "<NA>"
    } else {
      for (i in 1:(length(ftab) - 1)) {
        valstring <- paste0(valstring, ftab[i])
        if (i < length(ftab)) valstring <- paste0(valstring, "<br>")
      }
    }
  } else {
    valstring <- ""
  }
  return(valstring)
}

prc.value <- function(index, x, df.val, weights = NULL) {
  valstring <- ""
  # check for valid indices
  if (index <= ncol(x) && !is.null(df.val[[index]])) {
    # do we have weights?
    if (!is.null(weights))
      variab <- sjstats::weight(x[[index]], weights)
    else
      variab <- x[[index]]
    # create frequency table, but only get valid percentages
    ftab <- create.frq.df(variab, 20)$mydat$valid.prc
    # remove last value, which is a NA dummy
    if (length(ftab) == 1 && is.na(ftab)) {
      valstring <- "<NA>"
    } else {
      for (i in 1:(length(ftab) - 1)) {
        valstring <- paste0(valstring, sprintf("%.2f", ftab[i]))
        if (i < length(ftab)) valstring <- paste0(valstring, "<br>")
      }
    }
  } else {
    valstring <- ""
  }
  return(valstring)
}
  
  encoding <- get.encoding(encoding, x)
  if (!is.data.frame(x)) 
    stop("Parameter needs to be a data frame!", call. = FALSE)
  df.var <- sjlabelled::get_label(x)
  df.val <- sjlabelled::get_labels(x)
  colcnt <- ncol(x)
  id <- seq_len(colcnt)
  if (sort.by.name) 
    id <- id[order(colnames(x))]
  tag.table <- "table"
  tag.thead <- "thead"
  tag.tdata <- "tdata"
  tag.arc <- "arc"
  tag.caption <- "caption"
  tag.omit <- "omit"
  css.table <- "border-collapse:collapse; border:none;"
  css.thead <- "border-bottom:double; font-style:italic; font-weight:normal; padding:0.2cm; text-align:left; vertical-align:top;"
  css.tdata <- "padding:0.2cm; text-align:left; vertical-align:top;"
  css.arc <- "background-color:#eeeeee"
  css.caption <- "font-weight: bold; text-align:left;"
  css.omit <- "color:#999999;"
  if (!is.null(CSS)) {
    if (!is.null(CSS[["css.table"]])) 
      css.table <- ifelse(substring(CSS[["css.table"]], 
        1, 1) == "+", paste0(css.table, substring(CSS[["css.table"]], 
        2)), CSS[["css.table"]])
    if (!is.null(CSS[["css.thead"]])) 
      css.thead <- ifelse(substring(CSS[["css.thead"]], 
        1, 1) == "+", paste0(css.thead, substring(CSS[["css.thead"]], 
        2)), CSS[["css.thead"]])
    if (!is.null(CSS[["css.tdata"]])) 
      css.tdata <- ifelse(substring(CSS[["css.tdata"]], 
        1, 1) == "+", paste0(css.tdata, substring(CSS[["css.tdata"]], 
        2)), CSS[["css.tdata"]])
    if (!is.null(CSS[["css.arc"]])) 
      css.arc <- ifelse(substring(CSS[["css.arc"]], 1, 
        1) == "+", paste0(css.arc, substring(CSS[["css.arc"]], 
        2)), CSS[["css.arc"]])
    if (!is.null(CSS[["css.caption"]])) 
      css.caption <- ifelse(substring(CSS[["css.caption"]], 
        1, 1) == "+", paste0(css.caption, substring(CSS[["css.caption"]], 
        2)), CSS[["css.caption"]])
    if (!is.null(CSS[["css.omit"]])) 
      css.omit <- ifelse(substring(CSS[["css.omit"]], 
        1, 1) == "+", paste0(css.omit, substring(CSS[["css.omit"]], 
        2)), CSS[["css.omit"]])
  }
  page.style <- sprintf("<style>\nhtml, body { background-color: white; }\n%s { %s }\n.%s { %s }\n.%s { %s }\n.%s { %s }\n%s { %s }\n.%s { %s }\n</style>", 
    tag.table, css.table, tag.thead, css.thead, tag.tdata, 
    css.tdata, tag.arc, css.arc, tag.caption, css.caption, 
    tag.omit, css.omit)
  toWrite <- sprintf("<html>\n<head>\n<meta http-equiv=\"Content-type\" content=\"text/html;charset=%s\">\n%s\n</head>\n<body>\n", 
    encoding, page.style)
  page.content <- sprintf("<table>\n  <caption>Data frame: %s</caption>\n", 
    deparse(substitute(x)))
  page.content <- paste0(page.content, "  <tr>\n    ")
  if (show.id) 
    page.content <- paste0(page.content, "<th class=\"thead\">ID</th>")
  page.content <- paste0(page.content, "<th class=\"thead\">Name</th>")
  if (show.type) 
    page.content <- paste0(page.content, "<th class=\"thead\">Type</th>")
  page.content <- paste0(page.content, "<th class=\"thead\">Label</th>")
  if (show.na) 
    page.content <- paste0(page.content, "<th class=\"thead\">missings</th>")
  if (show.values) 
    page.content <- paste0(page.content, "<th class=\"thead\">Values</th>")
  if (show.labels) 
    page.content <- paste0(page.content, "<th class=\"thead\">Value Labels</th>")
  if (show.frq) 
    page.content <- paste0(page.content, "<th class=\"thead\">Freq.</th>")
  if (show.prc) 
    page.content <- paste0(page.content, "<th class=\"thead\">%</th>")
  if (show.wtd.frq) 
    page.content <- paste0(page.content, "<th class=\"thead\">weighted Freq.</th>")
  if (show.wtd.prc) 
    page.content <- paste0(page.content, "<th class=\"thead\">weighted %</th>")
  page.content <- paste0(page.content, "\n  </tr>\n")
  if (!hide.progress) 
    pb <- utils::txtProgressBar(min = 0, max = colcnt, style = 3)
  for (ccnt in seq_len(colcnt)) {
    index <- id[ccnt]
    arcstring <- ""
    if (altr.row.col) 
      arcstring <- ifelse(sjmisc::is_even(ccnt), " arc", 
        "")
    page.content <- paste0(page.content, "  <tr>\n")
    if (show.id) 
      page.content <- paste0(page.content, sprintf("    <td class=\"tdata%s\">%i</td>\n", 
        arcstring, index))
    if (!is.list(x[[index]]) && !is.null(sjlabelled::get_note(x[[index]]))) 
      td.title.tag <- sprintf(" title=\"%s\"", sjlabelled::get_note(x[[index]]))
    else td.title.tag <- ""
    page.content <- paste0(page.content, sprintf("    <td class=\"tdata%s\"%s>%s</td>\n", 
      arcstring, td.title.tag, colnames(x)[index]))
    if (show.type) {
      vartype <- sjmisc::var_type(x[[index]])
      page.content <- paste0(page.content, sprintf("    <td class=\"tdata%s\">%s</td>\n", 
        arcstring, vartype))
    }
    if (index <= length(df.var)) {
      varlab <- df.var[index]
      if (!is.null(wrap.labels)) {
        varlab <- sjmisc::word_wrap(varlab, wrap.labels, 
          "<br>")
      }
    }
    else {
      varlab <- "<NA>"
    }
    page.content <- paste0(page.content, sprintf("    <td class=\"tdata%s\">%s</td>\n", 
      arcstring, varlab))
    if (show.na) {
      if (is.list(x[[index]])) {
        page.content <- paste0(page.content, sprintf("    <td class=\"tdata%s\"><span class=\"omit\">&lt;list&gt;</span></td>\n", 
          arcstring))
      }
      else {
        page.content <- paste0(page.content, sprintf("    <td class=\"tdata%s\">%i (%.2f%%)</td>\n", 
          arcstring, sum(is.na(x[[index]]), na.rm = T), 
          100 * sum(is.na(x[[index]]), na.rm = T)/nrow(x)))
      }
    }
    if (is.numeric(x[[index]]) && !has_value_labels(x[[index]])) {
      if (show.values || show.labels) {
        valstring <- paste0(sprintf("%a", range(x[[index]], 
          na.rm = T)), collapse = "-")
        if (show.values && show.labels) {
          colsp <- " colspan=\"2\""
          valstring <- paste0("<em>range: ", valstring, 
            "</em>")
        }
        else {
          colsp <- ""
        }
        page.content <- paste0(page.content, sprintf("    <td class=\"tdata%s\"%s>%s</td>\n", 
          arcstring, colsp, valstring))
      }
    }
    else {
      if (show.values) {
        valstring <- ""
        if (index <= ncol(x)) {
          if (is.list(x[[index]])) {
            valstring <- "<span class=\"omit\">&lt;list&gt;</span>"
          }
          else {
            vals <- sjlabelled::get_values(x[[index]])
            if (!is.null(vals)) {
              loop <- na.omit(seq_len(length(vals))[1:max.len])
              for (i in loop) {
                valstring <- paste0(valstring, vals[i])
                if (i < length(vals)) 
                  valstring <- paste0(valstring, "<br>")
              }
              if (max.len < length(vals)) 
                valstring <- paste0(valstring, "<span class=\"omit\">&lt;...&gt;</span>")
            }
          }
        }
        else {
          valstring <- "<NA>"
        }
        page.content <- paste0(page.content, sprintf("    <td class=\"tdata%s\">%s</td>\n", 
          arcstring, valstring))
      }
      if (show.labels) {
        valstring <- ""
        if (index <= length(df.val)) {
          if (is.list(x[[index]])) {
            valstring <- "<span class=\"omit\">&lt;list&gt;</span>"
          }
          else {
            vals <- df.val[[index]]
            if (!is.null(vals)) 
              vals <- na.omit(vals)
            if (is.character(x[[index]]) && !is.null(vals) && 
              !sjmisc::is_empty(vals)) {
              if (show.string.values) 
                vals <- sort(vals)
              else vals <- "<span class=\"omit\" title =\"'show.string.values = TRUE' to show values.\">&lt;output omitted&gt;</span>"
            }
            if (!is.null(vals)) {
              loop <- na.omit(seq_len(length(vals))[1:max.len])
              for (i in loop) {
                valstring <- paste0(valstring, vals[i])
                if (i < length(vals)) 
                  valstring <- paste0(valstring, "<br>")
              }
              if (max.len < length(vals)) 
                valstring <- paste0(valstring, "<span class=\"omit\">&lt;... truncated&gt;</span>")
            }
          }
        }
        else {
          valstring <- "<NA>"
        }
        page.content <- paste0(page.content, sprintf("    <td class=\"tdata%s\">%s</td>\n", 
          arcstring, valstring))
      }
    }
    if (show.frq) {
      if (is.list(x[[index]])) 
        valstring <- "<span class=\"omit\">&lt;list&gt;</span>"
      else valstring <- frq.value(index, x, df.val)
      page.content <- paste0(page.content, sprintf("    <td class=\"tdata%s\">%s</td>\n", 
        arcstring, valstring))
    }
    if (show.prc) {
      if (is.list(x[[index]])) 
        valstring <- "<span class=\"omit\">&lt;list&gt;</span>"
      else valstring <- prc.value(index, x, df.val)
      page.content <- paste0(page.content, sprintf("    <td class=\"tdata%s\">%s</td>\n", 
        arcstring, valstring))
    }
    if (show.wtd.frq && !is.null(weight.by)) {
      if (is.list(x[[index]])) 
        valstring <- "<span class=\"omit\">&lt;list&gt;</span>"
      else valstring <- frq.value(index, x, df.val, weight.by)
      page.content <- paste0(page.content, sprintf("    <td class=\"tdata%s\">%s</td>\n", 
        arcstring, valstring))
    }
    if (show.prc && !is.null(weight.by)) {
      if (is.list(x[[index]])) 
        valstring <- "<span class=\"omit\">&lt;list&gt;</span>"
      else valstring <- prc.value(index, x, df.val, weight.by)
      page.content <- paste0(page.content, sprintf("    <td class=\"tdata%s\">%s</td>\n", 
        arcstring, valstring))
    }
    if (!hide.progress) 
      utils::setTxtProgressBar(pb, ccnt)
    page.content <- paste0(page.content, "  </tr>\n")
  }
  if (!hide.progress) 
    close(pb)
  page.content <- paste(page.content, "</table>", sep = "\n")
  toWrite <- paste0(toWrite, sprintf("%s\n</body></html>", 
    page.content))
  knitr <- page.content
  knitr <- gsub("class=", "style=", knitr, fixed = TRUE, useBytes = TRUE)
  knitr <- gsub("<table", sprintf("<table style=\"%s\"", css.table), 
    knitr, fixed = TRUE, useBytes = TRUE)
  knitr <- gsub(tag.tdata, css.tdata, knitr, fixed = TRUE, 
    useBytes = TRUE)
  knitr <- gsub(tag.thead, css.thead, knitr, fixed = TRUE, 
    useBytes = TRUE)
  knitr <- gsub(tag.arc, css.arc, knitr, fixed = TRUE, useBytes = TRUE)
  if (remove.spaces) {
    knitr <- sju.rmspc(knitr)
    toWrite <- sju.rmspc(toWrite)
    page.content <- sju.rmspc(page.content)
  }
  structure(class = c("sjTable", "view_df"), list(page.style = page.style, 
    page.content = page.content, output.complete = toWrite, 
    header = NULL, knitr = knitr, file = file, show = !no.output, 
    use.viewer = use.viewer))
}

view_df2(arab4, hide.progress = T)